The Saw-it-All-Along Effect: Demonstrations of Visual Hindsight Bias*
|
|
- Brenda Terry
- 5 years ago
- Views:
Transcription
1 The Saw-it-All-Along Effect: Demonstrations of Visual Hindsight Bias* Erin M. Harley University of California, Los Angeles Keri A. Carlsen and Geoffrey R. Loftus University of Washington Correspondence should be addressed to Erin M. Harley Department of Psychology, University of California, Los Angeles, Box , 1285 Franz Hall, Los Angeles, CA *The writing of this article and the laboratory described in it were supported by NIMH Grant MH41637 to G. Loftus. We thank Daniel Bernstein for valuable comments on previous versions of this manuscript. We also thank Chris Dessert, Eunice Poon, Sarah Hayes, and Erik Gust for assistance in data collection.
2 Abstract This article addresses whether a hindsight bias exists for visual perception tasks. In three experiments observers equipped with visual target information predicted the performance of a naïve self or peer. In phase 1 of all experiments, observers identified degraded images of celebrity faces as they resolved to full clarity. Following phase 1, Experiment-1 and 2 observers were given a surprise memory test in which they attempted to adjust the clarity of each face until it looked as it did when recognized in phase 1. Experiment-3 observers viewed phase-1 faces plus a new set of faces in full clarity at the outset, and then stopped the resolution process when they thought a naive peer would identify the face. In all experiments, observers overestimated identification performance of naïve observers. This visual hindsight bias was greater for more familiar faces those shown in both phases of the experiment and was not reduced following education and instructions to observers to avoid the bias. We conclude that once the content of a visual image is known, observers are unable to discount this information. As a result, they overestimate their ability, and the ability of others, to identify the image at a more degraded state. We propose a fluency misattribution theory to account for the bias. Implications for medical malpractice litigation and eyewitness testimony are discussed.
3 HARLEY, CARLSEN, & LOFTUS 1 Imagine the following scenario: A patient at a local hospital undergoes routine chest radiography at time 1, and Radiologist 1 interprets the films as normal. Three years later, the patient experiences discomfort and undergoes chest radiography a second time, whereupon Radiologist 2 discovers a large tumor. Despite treatment, the patient dies. The patient s family files a medical malpractice lawsuit against Radiologist 1, claiming the tumor should have been detected at time 1. Radiologist 2, testifying on the family s behalf, views the original time-1 radiographs and, seeing the tumor missed by Radiologist 1, claims that it was visible at time 1. Such a scenario is typical; the vast majority of medical malpractice litigation in Radiology over the past 20 years has involved diagnostic, perceptual, or decision-making errors (Berlin, 1996a, 1996b; Berlin & Hendrix, 1998). In such cases, the defendant radiologist being accused of negligence for a decision made prospectively (without the benefit of outcome knowledge) is often judged by radiology experts who have full knowledge of what future radiographs revealed (Berlin, 2000). Are such experts biased by this knowledge; i.e., can a radiologist who views a tumor at time 2 make an appropriate prediction about whether another radiologist should have detected the tumor when it was smaller and less visible at time 1? This is a question of visual hindsight, which is akin to verbal hindsight reported in the past by many investigators. First reported by Fischhoff (1975), hindsight bias, or the knew-it-all-along effect (Wood, 1978), is the tendency for individuals with outcome knowledge to claim that they would have estimated a higher probability of occurrence for the reported outcome than was estimated in foresight. In other words, it is the after-the-fact feeling that some outcome was very likely to happen, or was predictable, even though it was not predicted to happen beforehand. Hindsight bias is thought to result from cognitive reconstruction processes that occur after outcome information is received (for a review see Hawkins & Hastie, 1990). Judges reanalyze the event so that the beginning and the middle connect causally to the end. During this process evidence consistent with the reported outcome is elaborated and evidence inconsistent with the outcome is minimized or discounted. The result of this rejudgment process is that the given outcome seems inevitable, or at least more plausible than alternative outcomes. The verbal hindsight bias effect has been shown to be quite robust, occurring in both within- and between-subject designs, in spite of explicit instructions to subjects to avoid the bias, and across a range of time intervals between initial judgments, outcome feedback, and second judgments (see Christensen- Szalanski & Willham, 1991; Hawkins & Hastie, 1990; Fischhoff, 1982; for reviews). The bias is sensitive to task difficulty it is greater for difficult than easy items, and greater for events initially judged to be
4 HARLEY, CARLSEN, & LOFTUS 2 least plausible (Fischhoff, 1977; Wood, 1978; Arkes, Wortmann, Saville, & Harkness, 1981) and is greater when evidence supporting the given outcome is more easily brought to mind (Sanna, Schwarz, & Small, 2002). Hindsight bias has been studied extensively in the cognitive domain, with broadly ranging types of events, e.g., outcomes of historical events, psychiatric cases, scientific experiments, consumer purchases, sporting events, economic decisions, election outcomes, medical and legal cases, and answers to almanac trivia questions. In contrast, there has been remarkably little research on the kind of visual hindsight bias exemplified by the radiology example sketched earlier. The main purpose of this article is to begin to correct this deficit. We first describe how hindsight bias might apply to visual perception, and in this context, introduce the broader topic of metaperception people s insights into their perceptual abilities. Next, we present two experiments that show evidence of visual hindsight bias, and propose a theory of fluency misattribution to account for the bias. Finally, we present a third experiment in which a prediction of the fluency-misattribution theory is tested and confirmed. Visual Hindsight Bias To illustrate how hindsight bias might apply to visual perception we return to the medical scenario described earlier. In the field of Radiology, it is common in medical malpractice litigation for a physician with outcome knowledge to judge whether another physician, who did not have the benefit of such knowledge, should have detected an abnormality on a medical image (Berlin, 2000). When making this judgment, it may upon initial inspection seem reasonable for the judging physician to simply assess his or her own ability to see the missed abnormality. However, the physician cannot use this method if the missed abnormality is more easily detected in hindsight. There is evidence to suggest that medical abnormalities are in fact more visible when the physician has outcome knowledge. A striking example of tumors only evident in hindsight comes from Muhm, Miller, Fontana, Sanderson, and Uhlenhopp (1983), who conducted a screening program at the Mayo Clinic for men at high risk of lung cancer. The 4,618 members of the study group obtained chest radiographs every four months, each of which was read by 2-3 radiologists or chest physicians. Over the course of the 6-year study, 92 tumors were detected in the study group. Of these, 75 (82%) were, as the authors termed it, visible in retrospect (p. 611). This means that when the physicians looked at the previous sets of radiographs, in 82% of the cases they found they could detect the tumor on at least one,
5 HARLEY, CARLSEN, & LOFTUS 3 and often multiple, prior radiographs (ranging from 4 to 53 months prior) that had initially been interpreted as normal. When speculating as to why the 75 tumors visible in hindsight were not detected on initial readings, Muhm et al. (1983) made the following comment: The fact that some of the cancers had been overlooked was usually due to perceptual errors [italics added] by the observers (p. 612). Does it make sense that 2-3 experienced radiologists made errors on 82% of the tumor-containing radiographs? We argue that the error may well lie not in the physicians initial interpretations of the radiographs, but rather in their retrospective interpretations of the radiographs. The overlooked tumors became visible in hindsight not because they were more detectable to begin with, but because the physicians had the benefit of outcome knowledge. Multiple factors contribute to allow a radiologist with outcome knowledge, or hindsight, to detect abnormalities that previously could not be seen. Since Bartlett s (1932) pioneering work, and Neisser s landmark Cognitive Psychology (1967), visual perception has been construed as an active and creative process involving far more than a direct translation of images projected onto the retina. What we perceive is influenced by top-down processes that include prior knowledge, expectations, context, and a large number of assumptions about how objects in the world behave (see, e.g., Palmer, 1999). The radiologists who initially read the films in the Mayo Clinic s screening program had relatively low expectations for finding a tumor 1, and if a tumor were present, they had no knowledge of what type of tumor to search for, or where the tumor might be located. In contrast, the post-detection viewing conditions were quite different. During retrospective analysis of the radiographs the physicians had high expectations for finding a tumor, had full knowledge of both tumor type and tumor location, and because the outcome was known, no longer had to be cautious about avoiding false positives. These factors combined to make visible many of the abnormalities that had originally gone undetected. There is some research indicating that observers who know what they are looking for will have higher detection rates in a difficult visual search task. Bruner and Potter (1964) reported that observers claim to be able to perceive a visual target at a more degraded state when viewed in a clear-to-blurry progression than in a blurry-to-clear progression. Given that target information can improve an observer s ability to detect or identify a visual target, an observer asked to estimate the performance of a naïve peer is faced with a difficult challenge. The observer cannot simply assess his or her current detection ability it has 1 Even though the patients were at high risk for lung cancer, over the 6-year screening only 92 tumors were detected in a group of 4,618 patients (roughly 2% of the sample).
6 HARLEY, CARLSEN, & LOFTUS 4 been enhanced by outcome information but instead must discount target information, and imagine what a naïve observer would perceive. Metaperception When asked to estimate the visual performance of a naïve self or peer, an observer must attempt to gain insight into the strengths and limitations of their perceptual abilities, as well as how these abilities are affected by experience and knowledge. We will term this insight metaperception. Although a large amount of research has been dedicated to the investigation of metacognition, and most specifically, metamemory (see Metcalfe & Shimamura, 1994; Mazzoni & Nelson, 1998; Chambres, Izaute & Marescaux, 2002, for reviews), very few researchers have explored metaperception. Recent work by Levin, Momen, Drivdahl, & Simons (2000), and Levin (2002) has investigated metaperception in relation to change blindness, the counterintuitive finding that observers commonly fail to detect large visual changes in their environment (see, e.g., Blackmore, Brelstaff, Nelson, & Troscianko, 1995; Henderson, 1997; Pashler, 1988; Phillips, 1974; Rensink, O'Regan, & Clark, 1997; for a review see Simons, 2000). The existence of change blindness suggests that observers do not retain many visual details in memory from one view to the next, and that focusing attention on the changing item is critical for successful change detection. Most relevant to the present work is that change blindness is a startling finding in that it contradicts people s intuitions about their perceptual abilities. Prior to experience with change-detection tasks, most observers wrongly assume that they will have no trouble detecting changes in visual scenes (Levin, Drivdahl, Momen, & Beck, 2002). For example, Levin et al (2002) found that 90% of participants believed they would notice a scarf disappear on an actor across a movie cut that 0% of observers in the original experiment noticed. This metacognitive error has been termed change blindness blindness, and is evidence that under some conditions naïve observers have grossly inaccurate insights into their own and others perceptual abilities. This conflict between actual visual performance and estimated visual performance warrants further investigation of metaperception. Specifically, under what conditions will observers make inaccurate estimates of their perceptual abilities? Judging in hindsight may be one such case. Only two studies that we are aware of have examined hindsight bias in the visual domain. Winman, Juslin, & Bjorkman (1998) found a reverse visual hindsight bias, i.e., an underestimate of performance, when observers with outcome information estimated prior, naïve performance in a two-alternative-forcedchoice line-length discrimination task. In contrast, Harley, Carlsen, & Loftus (2001) found that observers
7 HARLEY, CARLSEN, & LOFTUS 5 show positive hindsight bias when predicting the performance of their peers in a digit-identification task, but only when the task was most difficult. The conflicting results of these two studies suggest that we cannot assume a priori that hindsight bias will exist for all perceptual tasks, nor can we assume that if it does it will be a positive bias, like the traditional verbal hindsight bias, rather than a negative bias, like the one found by Winman et al (1998). Another reason we cannot assume a priori that visual hindsight bias will mirror verbal hindsight bias is that confidence in foresight knowledge judgments appears to trend differently in the intellectual and sensory domains. People tend to be overconfident when making intellectual judgments, e.g., Who was the third president of the United States, but underconfident when making sensory judgments, e.g., What color are your colleague s eyes? (e.g., Adams, 1957; Bjorkman, Juslin, & Winman, 1993; Dawes, 1980; Keren, 1988; Olsson & Winman, 1996; Winman and Juslin, 1993). If hindsight confidence ratings are based on the same information as foresight confidence ratings Winman, Juslin, & Bjorkman (1998) suggest that the two are both based on an assessment of task difficulty then a reverse hindsight bias may prove more prevalent in the visual domain. If, on the other hand, hindsight bias is a general phenomenon that affects decisions made in multiple domains, including visual, then we expect to find evidence for a positive visual hindsight bias, i.e., an overestimation of naïve performance after target identity is known. In a series of experiments reported here we tested the hypothesis that hindsight bias is a general phenomenon an overestimation of foresight knowledge following the receipt of outcome knowledge that is not restricted to the intellectual domain, but occurs in the visual domain as well. To test this hypothesis we examined whether an observer with knowledge of target identity, akin to outcome knowledge, could accurately predict the level of visual degradation at which a naïve self or peer would be able to identify a celebrity face. In all experiments observers exhibited visual hindsight bias. The bias was exhibited for judgments made about both self and others, as well as in spite of education and warnings to avoid the bias. We propose a fluency-misattribution theory to account for visual hindsight bias and provide confirmatory evidence of a prediction of the theory in Experiment 3. Experiment 1 In Experiment 1, observers identified degraded pictures of celebrity faces as they gradually clarified. Later, in a surprise memory test, observers recalled the degree of degradation present at time of original identification.
8 HARLEY, CARLSEN, & LOFTUS 6 Method Observers Forty-two University of Washington undergraduates, all with normal or corrected-to-normal vision, participated in exchange for course credit. Apparatus Data-collection took place in a room equipped with four Macintosh emac computers each equipped with a G4 processor and a 17" monitor allowing for up to four observers to participate in each datacollection session. Curtains were hung between computers to prevent observers from viewing other monitors during the session. Each data collection session lasted approximately 30 minutes. The experiment was written and executed in MATLAB using the Psychophysics Toolbox extensions (Brainard, 1997; Pelli, 1997). Stimuli Stimuli were 36 grayscale pictures of celebrity faces. The set included well-known actors, musicians, politicians, and sports figures, e.g., Jerry Seinfeld, Harrison Ford, Madonna, Hillary Clinton, and Michael Jordon (see Figure 1). Each face measured 500 pixels from the bottom of the chin to the top of the head, and subtended a visual angle of about 21.8 vertically. The display monitor s background luminance measured 2.94 cd/m 2. Thirty successively more blurred versions of each face were created. Each blurred version was accomplished by Fourier-transforming the image from pixel space into spatial-frequency space, multiplying the resulting frequency amplitude spectrum by a Gaussian low-pass filter and inverse-fourier transforming the result back into pixel space. The low-pass filter was designed such that for each blur level it passed frequencies perfectly, i.e., had a value of 1.0 up to some value f 0 cycles per face height, and then fell parabolically, reaching zero at the cutoff frequency value of f 1 =f 0 *3 cycles per face height. As f 0 and f 1 are made smaller, the filter cuts off more spatial frequencies, and the resulting face becomes blurrier (see Loftus, 2001 for additional explanation). See Figure 1, row 2, for samples of a filtered face. In each of the three experiments reported here, f 1, the filter cutoff frequency, was used as the dependent variable a measure of the degree of blur present in the image when the observer was able to (or believed they would be able to) identify the celebrity. We reemphasize that a blurrier picture is implied by a smaller f 1 value.
9 HARLEY, CARLSEN, & LOFTUS 7 Jerry Seinfeld Harrison Ford Madonna Hillary Clinton Michael Jordan f 1 = 5.16 f 1 = 9.90 f 1 = f 1 = unfiltered Figure 1. Samples of celebrity faces used in Experiments 1-3. Row two contains 4 of the 30 lowpass filtered versions of Jennifer Aniston, with corresponding f 1 (cycles per face height) filter cutoff values. Design and Procedure Phase 1 of the experiment, baseline identification, was a simple identification test. For each face, the 30 blur images were displayed, in order, from most to least blurred, at a rate of 500 ms per image. To the observer it appeared as if the celebrity s face was slowly clarifying over time. Observers were read aloud the following instructions: Each celebrity will start very blurry and slowly become clear. Press the space bar as soon as you recognize who the celebrity is. After you press the space bar type in this guess and hit RETURN. After you enter your guess the picture will continue to become clear. If your guess changes hit the space bar again and type in a new guess. You can guess as many times as you wish until you are certain you have identified the celebrity correctly. If you get all the way to the clearest image and you don t know who the celebrity is just type in don t know or?. Observers were allowed to identify a celebrity in a number of ways, including, any portion of the celebrity s name, the name of a character they play on TV, the name of a movie in which they have starred, etc. Anything that would indicate to the investigators that the celebrity was recognized was scored
10 HARLEY, CARLSEN, & LOFTUS 8 as an accurate response. At the completion of each trial the observer was asked to verify his or her final identity guess. For each celebrity face all 30 blur images were displayed regardless of whether or when the face was identified. This was done to equate as best as possible the total time an observer viewed each face. The order in which the 36 celebrities appeared was randomized for each observer. In phase 2 of the experiment, observers were given a surprise memory test. The same celebrities were shown again in a different, random order. At the start of each trial the face was shown in its most degraded form and the observer s response from phase 1 was printed on the screen below the face so that the celebrity s identity would be known despite beginning in a degraded state. Observers were read aloud the following instructions: Now you are going to perform a memory test. Use the arrow keys to adjust the blurriness of the celebrity until it matches what the celebrity looked like when you correctly identified him/her in the first half of the experiment. Observers were allowed to range back and forth among the 30 filters until they were satisfied with their decisions; no time limits were imposed. Observers completed two practice trials prior to each of the two tasks: baseline-id and memory test. Celebrities shown in practice trials did not appear in the experiment proper. Results and Discussion Evidence for visual hindsight bias was found in Experiment 1. When asked to recall how blurry each face looked when identified in Phase 1 of the experiment, observers systematically overestimated the degree of blur. Only trials for which an observer correctly identified the celebrity during baseline identification were included in the analysis. The mean baseline-id point was f 1 = 25.68, whereas the mean estimated identification point from the memory test was f 1 = 20.73, implying that the observers remembered identifying the celerity in a blurrier state (f 1 = 20.73) than was actually the case (f 1 = 25.68). The hindsight ratio (HR) can be quantified as the ratio of these two numbers: baseline-id f 1 over memorytest f 1. To the extent that observers show hindsight bias, HR will be greater than 1, as is the case here: HR = 1.28 ± (see Figure 2A). The bias was not driven by a few outlying observers; 37 of the 42 observers (88%) showed visual hindsight bias. As mentioned in the Introduction, there is evidence that verbal hindsight bias is greater for difficult than easy items (Fischhoff, 1977; Wood, 1978). To assess the influence of task difficulty on the 2 The notation x ± y refers to a mean plus or minus a 95% confidence interval.
11 HARLEY, CARLSEN, & LOFTUS 9 A 2.6 Experiment-1 Data B 2.6 Experiment-2 Data Hindsight Ratio Hindsight Ratio All Faces Difficulty Quartile (less difficult ==> more difficult) 0.6 All Faces Difficulty Quartile (less difficult ==> more difficult) Figure 2. A: Experiment-1 data (n = 42). Average hindsight ratios (baseline-id f 1 divided by memory-test f 1 ) are plotted for all faces, and for faces divided into four difficulty quartiles. Error bars represent 95% confidence intervals. B: Experiment-2 data (n = 54). magnitude of visual hindsight bias, celebrities were divided into four difficulty quartiles based on baseline-id point. Faces with lower f 1 identification values, i.e., identified at a more degraded state, were considered easier than those with higher f 1 identification values, i.e., identified at a less degraded state. This was done separately for each observer, and the means for each quartile were then averaged across observers. Hindsight ratios for the four difficulty quartiles are shown in Figure 2A. No bias was found for the easiest faces, HR = 0.99 ± For the remaining three difficulty quartiles, hindsight ratios increase as difficulty increases, ranging from a small bias for faces in difficulty quartile 2, HR = 1.18 ± 0.06, to a large bias for the most difficult faces, HR = 1.72 ± Experiment 2 In the verbal domain, many researchers have tried to eliminate or at least reduce the hindsight bias effect, and most attempts to do this have been unsuccessful. Fischhoff (1977) found that neither educating participants about hindsight bias nor warning them to do everything they could to avoid the bias were successful in reducing the effect. In Experiment 2 we used a similar approach to test whether visual hindsight bias is cognitively impenetrable. A mental function is said to be cognitively impenetrable if it cannot be influenced by purely cognitive factors such as goals, beliefs, and inferences (Pylyshyn, 1980).
12 HARLEY, CARLSEN, & LOFTUS 10 Experiment 2 was a replication of Experiment 1 with one addition: Immediately prior to the memory test, observers were educated about visual hindsight bias, and were warned to avoid it and perform as accurately as possible. Method Observers Fifty-four University of Washington undergraduates, all with normal or corrected-to-normal vision, participated in exchange for course credit. Apparatus and Stimuli All equipment and stimuli were identical to that used in Experiment 1. Design and Procedure The design of Experiment 2 was identical to that of Experiment 1 with one difference. Following the baseline-identification task and instructions on how to perform the memory test, observers were read aloud the following: When remembering how blurry each celebrity was when you first recognized him or her, I would like you to be aware of hindsight bias. Hindsight bias is when someone who knows the outcome of an event thinks they would have predicted that outcome before it happened. In previous versions of this experiment, your peers tended to believe that they recognized celebrities earlier than they did in phase 1 of the experiment. We believe that already knowing who the celebrity is before viewing the clarification is what causes this effect. The result is that your peers think they recognized celebrities earlier, at a blurrier point, than they actually did. Please try to avoid this bias and be as accurate as possible when performing the memory test. Results and Discussion The instructions and warning given to observers in Experiment 2 did not reduce the size of the hindsight bias. As was the case with Experiment-1 observers, Experiment-2 observers systematically overestimated the degree of blur when asked to recall baseline-id performance. Data, averaged across 54 observers, are shown in Figure 2B. Only trials for which an observer correctly identified the celebrity during the baseline-id task were included in the analysis. The mean baseline-id point was f 1 = 29.13, whereas the mean estimated identification point from the memory test was f 1 = Recall that the hindsight ratio (HR) can be quantified as the ratio of these two numbers: baseline-id f 1 over memory-test f 1 ; HR = 1.31 ± 0.10.
13 HARLEY, CARLSEN, & LOFTUS 11 To assess the influence of task difficulty on the magnitude of the bias, celebrities were divided into four difficulty quartiles using the procedure used for the Experiment-1 data. The difficulty effect found in Experiment 1 was replicated in Experiment 2 (see Figure 2B). Identical to the pattern observed in the Experiment-1 data, no bias was found for the easiest faces, HR = 0.92 ± 0.05, and for the remaining three difficulty quartiles, hindsight ratios increase as difficulty increases, ranging from a small bias for faces in difficulty quartile 2, HR = 1.10 ± 0.06, to a large bias for the most difficult faces, HR = 2.09 ± Note that there was no reduction in the hindsight bias effect observed in Experiment 2 when compared to the original Experiment-1 data. Educating observers about hindsight bias, and warning them to avoid the bias and perform as accurately as possible were not effective in reducing bias. Experiment-2 data suggest that visual hindsight bias is similar to verbal hindsight bias in that they are both automatic, unconscious processes that observers are not easily able to control. A Fluency-Misattribution Theory of Visual Hindsight Bias When an observer views a visual stimulus, it is processed with some degree of perceptual fluency that reflects processing speed, effort, and accuracy. Stimulus variables such as clarity, long exposure duration, familiarity and semantic relatedness, can serve to enhance perceptual fluency. Jacoby and Whitehouse (1989) demonstrated that if participants are unaware of why fluency has been enhanced, they might misattribute the fluency to something else, e.g., in their study, recent prior exposure to the stimulus. Since this work, research has shown that enhanced fluency can be misattributed to numerous factors in addition to prior exposure, such as, stimulus duration, clarity, truth, and liking (for a review see Winkielman, Schwarz, Reber, & Fazendeiro, 2003). We propose that visual hindsight bias results from misattribution of enhanced perceptual fluency following the receipt of target-identity information. In the present experiments, processing of a degraded face after target identity is known (during the memory test) is enhanced compared to when it is not known (during the baseline-id task). Observers must ignore the enhanced fluency if they are to accurately estimate when a naïve self or peer will make a correct identification. If observers fail to discount the enhanced fluency, or do not discount enough, they may mistakenly misattribute some or all of it to the predictability of the given outcome resulting in an overestimation of naïve performance, i.e., hindsight bias (see Bernstein, Whittlesea & Loftus, 2002; Whittlesea & Williams, 2000; for similar accounts).
14 HARLEY, CARLSEN, & LOFTUS 12 Experiment 3 The fluency-misattribution theory makes the following prediction: The more fluently a target is processed following the receipt of identity information, the larger should be the size of the hindsight effect. Experiment 3 was designed to test this prediction. Phase 1 was identical to the baseline-id task used in Experiments 1 and 2: Observers viewed celebrity faces as they clarified over time, and stopped the process when identification of the face was possible. Phase 2 differed from Experiments 1 and 2. Instead of a memory test, Experiment 3 used what is referred to as a hypothetical hindsight design in which observers with outcome information predict the performance of a naïve peer. Briefly, in the hindsight task observers viewed an outcome stimulus an unfiltered version of the face followed by the same clarification process used during the baseline-id task. Observers stopped the clarification when a naïve peer, someone who did not see the outcome stimulus, would be able to identify the celebrity. Critically, two sets of faces were shown in phase 2: the faces shown during phase 1 (old faces), and new, previously unseen faces. We expect observers to show hindsight bias for both types of faces because processing fluency during the hindsight task will be enhanced as a result of viewing the outcome stimuli. However, we expect a larger bias for old faces compared to new because processing fluency for old faces will receive additional enhancement from viewing those faces during the baseline-id task. Experiment 3 also allows for a test of the cognitive reconstruction theories proposed to account for verbal hindsight bias. Note that for the new faces shown in the hindsight task observers will view the outcome stimulus prior to any exposure to the degraded face. Cognitive reconstruction theories cannot account for bias in an outcome-first presentation because at the time outcome information is received there is no prior evidence to be reworked or rejudged. If hindsight bias is observed for the new faces, rejudgment processes cannot be necessary to produce visual hindsight bias. Method Observers Fifty-three University of Washington undergraduates, all with normal or corrected-to-normal vision, participated in exchange for course credit. Apparatus and Stimuli All equipment and stimuli were identical to those used in Experiment 1.
15 HARLEY, CARLSEN, & LOFTUS 13 Design and Procedure Each observer participated in a baseline-identification task followed by a hindsight task. The baseline-identification task was identical to that used in Experiments 1 and 2. Briefly, each face progressed from highly degraded to full clarity over 15 seconds, and observers stopped the resolution process as soon as they recognized the face. Prior to the completing the hindsight task observers were read aloud the following instructions: In the second phase of the experiment you are going to see the same celebrities again plus some new ones. Instead of indicating at what point you recognize the celebrity, your task this time will be to estimate at what point one of your peers would recognize the celebrity. So that you know the correct answer we will show you a clear version of the celebrity at the beginning of each trial. If you know who they are type that in. If you do not know who the celebrity is, type in a question mark and it will skip to the next trial. After you identify the clear picture of the celebrity, the face will go blurry and slowly resolve- just like in phase 1. Now imagine that a same-age peer is seeing the face for the first time, meaning they did not see the clear picture first. Press the space bar when you think your peer would recognize who the person is. When the observer stopped the clarification process, the following question appeared on the screen, Is this the point at which you think your peer would recognize this person? If the observer entered y for yes, the trial ended. If the observer answered n for no, the clarification process continued. Observers were allowed to stop the process as many times as necessary until they were satisfied with the degree of blur present in the image. Half, 18 of 36, of the celebrities were shown in the baseline-id task. The 18 baseline faces plus 18 new faces were mixed and shown in the hindsight task. The choice of which celebrities were shown twice (i.e., shown in both the baseline-id and hindsight tasks), and the order of the 64 celebrities across trials, were counterbalanced across observers. Observers completed two practice trials prior to each of the two tasks. Celebrities shown in practice trials did not appear in the experiment proper. Results and Discussion Experiment-3 data are shown in Figure 3. Only trials for which an observer could correctly identify the celebrity were included in the analysis. Recall that the average f 1 value for the baseline-id condition represents the average degree of blur present at time of correct identification, whereas average f 1 values for the two hindsight conditions represent the degree of blur present when observers predicted a naïve peer would make an accurate identification. As predicted by the fluency-misattribution theory, average f 1
16 HARLEY, CARLSEN, & LOFTUS 14 identification point was greatest in the baseline-id condition (M = 29.8), followed by the new faces in hindsight (M = 21.4), followed by old faces in hindsight (M = 18.4). Two orthogonal planned comparisons (C1 and C2) were tested to compare the three conditions: Baseline-ID performance was compared to the two hindsight conditions (C1), and the two hindsight conditions were compared to one another (C2). For each observer a ratio of the weighted conditions was computed to test each comparison (sum of positive-weighted conditions over sum of negative-weighted conditions), and the mean of the ratios was computed. Compared to baseline-id, observers demonstrated hindsight bias for both the old and new faces, M C1 = 1.45 ± Additionally, as predicted by the fluency-misattribution theory, the hindsight effect was larger for old than for new faces, M C2 = 1.08 ± f <== more blurry 1 less blurry ==> Baseline Identification New Faces Old Faces Hindsight Conditions Figure 3. Experiment-3 data (n=53). Performance in the three conditions is plotted as a function of f 1, filter cutoff frequency expressed as cycles per face height. Error bars represent standard errors.
17 HARLEY, CARLSEN, & LOFTUS 15 General Discussion and Implications Data from three experiments provide evidence for visual hindsight bias; observers operating with the benefit of target identity information may be biased when asked to estimate the performance of a naïve peer or self. It was found for judgments made about both self (Experiments 1 and 2) and others (Experiment 3), and despite education and explicit instructions to observers to avoid the bias (Experiment 2). The bias was larger when post-outcome processing was more fluent, and when pre-outcome identification was more difficult. We now discuss these last two findings in more detail. Fluency-Misattribution Theory We have proposed a theory of fluency misattribution to account for visual hindsight bias. Exposure to outcome information increases the perceptual fluency with which a degraded image is processed. This fluency must be discounted if an observer is to accurately predict identification performance of a naïve self or peer. If the observer fails to fully discount the enhanced fluency, it is misattributed to the predictability of the outcome. The theory predicts larger hindsight bias for targets processed more fluently following the receipt of identity information. This prediction was confirmed in Experiment 3 in which a larger bias was found for old faces those shown both in the baseline-id task and the hindsight task compared to new faces shown only in hindsight. Jacoby and Whitehouse (1989) found that participants who are aware of why fluency has been enhanced were able to discount the fluency. Only when participants were made unaware of the source of the enhanced fluency did they misattribute it to something else. In the experiments reported here, observers may have had some awareness that processing fluency was enhanced via learning the identities of the celebrities during the baseline-identification task. Although the data clearly indicate that observers did not successfully discount all of the enhanced fluency, it is possible that observers discounted some of the enhanced fluency. If the source of increased fluency were made more obscure, a larger bias might be found. In fact, decreased awareness of the source of enhanced fluency may have contributed to the larger bias observed for old faces compared to new faces in Experiment 3. If hindsight bias is a general metacognitive error to which all modalities are vulnerable, as we believe it is, then the fluency-misattribution theory should account for verbal hindsight bias in addition to visual hindsight bias. There is some evidence that this may be the case. Using different terminology, Sanna, Schwarz, & Small (2002) suggest that fluency plays a large role in verbal hindsight bias. The authors demonstrated that hindsight bias is reduced when evidence supporting alternative outcomes is easy to
18 HARLEY, CARLSEN, & LOFTUS 16 bring to mind, but it is increased when such evidence is difficult to bring to mind. They term the ease or difficulty with which these thoughts are brought to mind, subjective accessibility, which we argue, is really just another way of describing fluency. The outcome, and evidence for it, that is processed more fluently is seen as more likely and more predictable. In the case of verbal hindsight bias, fluency is conceptual rather than perceptual. This work suggests that a fluency-misattribution theory may account for verbal hindsight bias as well as visual. Further studies are warranted to examine this possibility. Task Difficulty As mentioned in the Introduction, there is evidence that verbal hindsight bias is greater for difficult than easy items, and greater for events initially judged to be least plausible (Fischhoff, 1977; Wood, 1978; Arkes, Wortmann, Saville, & Harkness, 1981). In the visual domain, Harley, Carlsen, and Loftus (2001) found that visual hindsight bias only occurred for the most difficult-to-detect targets. The results reported here are consistent with these findings. For Experiments 1 and 2, no hindsight effect was observed for the easiest faces, and the effect then increased monotonically as the degree of foresight identification difficulty increased. The influence of task difficulty on the size of visual hindsight bias can be accounted for by the fluency-misattribution theory if we assume that outcome information is more beneficial to the processing of difficult targets. To illustrate, prior to the receipt of target identity information, some degraded targets will be more difficult to identify than others. Following the receipt of target identity information, all of the images become identifiable at a more degraded state. If target identity information is more beneficial for an item that was originally difficult to identify compared to one that was not difficult, then the discrepancy between baseline processing fluency and hindsight processing fluency will be larger for more difficult items. The greater this discrepancy, the more fluency an observer must discount to make an accurate prediction about a naïve observer s ability. As evidenced by Experiment-1 and Experiment-2 data, observers do adjust appropriately for the easiest faces, those for which the outcome information presumably provides the least benefit, but the adjustment falls shorter and shorter as difficulty increases. One could argue that the difficulty effect observed in Experiments 1 and 2 resulted from observers failure to stop the clarification process at the point of recognition due to an inability to remember a celebrity s name. Recall that faces were divided into difficulty quartiles based on the point at which they were identified during the baseline-id task; faces identified later, i.e., at a clearer state, were categorized as more difficult than those identified earlier. It is possible that faces identified later (coded as more difficult) were faces for which observers recognized the celebrity but had trouble recalling a name or
19 HARLEY, CARLSEN, & LOFTUS 17 other identifying remark. In such cases, observers may have let the clarification process continue until a name could be recalled. Although this account is plausible, we have two reasons to believe that observers were not performing the task in this manner. First, observers were instructed to, press the space bar as soon as you recognize who the celebrity is. They were also told that if their guess changed they could stop the clarification process again, and that there would be no penalty for guessing multiple times. Second, in watching observers perform the baseline-id task they appeared to stop the clarification process as soon as the face was recognizable, and then wait until a name or identifying remark could be generated before continuing. Given, however, that our second piece of evidence is anecdotal, and that we cannot know if all observers followed the directions, the alternative account of the difficulty effect cannot be ruled out. Legal Implications Visual hindsight bias has a number of important legal implications. We return once again to the medical malpractice scenario described at the outset of the paper. When asked to estimate whether Radiologist 1 should have detected a tumor at time 1, Radiologist 2 should proceed with extreme caution. Not only will the tumor be more visible to an observer operating with the benefit of outcome knowledge (e.g., Muhm, et al., 1983), but also, based on the results reported here, it is highly likely Radiologist 2 will not fully discount this benefit, and as a result, will overestimate the detection ability of a naïve observer. Physicians judging the visibility of missed tumors are not the only legal players who may be susceptible to visual hindsight bias. Eyewitnesses to a crime are often asked to judge their perceptual abilities, and in doing so, may overestimate their ability to detect or identify a visual target under poor viewing conditions. For example, if Warren Witness views someone fleeing from the scene of a crime 30 feet away, on a dark and rainy night, his memory of that person s face is likely to be poor. If Warren is later shown a clear picture of Sam Suspect in a police photo lineup (akin to outcome knowledge), Warren may integrate the clear picture of Sam with the original degraded image he has stored in memory. The result will be a cleaned up memory representation that more closely resembles Sam Suspect. This is clearly problematic, as Sam may or may not be the criminal Warren saw, but the trouble does not end here. When asked to testify against Sam Suspect, Warren Witness may overestimate his ability to have identified Sam under the original, poor viewing conditions. The notion that feedback can influence an eyewitness s estimate of original viewing conditions is not new. Verbal confirmatory post-identification feedback has been shown not only to inflate witnesses confidence in the accuracy of their identification, but also to distort their estimates of the original
20 HARLEY, CARLSEN, & LOFTUS 18 witnessing conditions (Wells & Bradfield, 1998; Bradfield, Wells, & Olson, 2002). Goodness of view, speed of identification, amount of attention paid to the suspect s face, and clarity of memory for the suspect, are given inflated ratings by eyewitnesses who received confirmatory feedback following identification of a suspect. The data reported here add to the mounting evidence that outcome information, be it verbal or visual, can distort an eyewitness s beliefs about the original viewing conditions. Concluding Remarks We have provided evidence that, under certain conditions, observers operating with the benefit of hindsight do not have accurate insights into the strengths and limitations of their perceptual abilities. This is evidenced by their failure to accurately predict the performance of a naïve peer or self in visual identification tasks. Like verbal hindsight bias, visual hindsight bias appears to be moderated by task difficulty, with a greater bias occurring for more difficult (i.e., ambiguous) images, and appears to be cognitively impenetrable. A fluency-misattribution theory was proposed to account for the bias. The theory posits that exposure to target identity information results in enhanced processing fluency of the degraded image. When asked to judge the performance of a naive observer, the increased fluency is not fully discounted, and instead may be misattributed to the predictability of the given outcome. The theory predicts a larger hindsight effect for more fluently processed targets. This prediction was confirmed in Experiment 3 in which a larger bias was observed for targets made more familiar via repeated exposures.
21 HARLEY, CARLSEN, & LOFTUS 19 References Adams, J. K. (1957). A confidence scale defined in terms of expected percentages. American Journal of Psychology, 70, Arkes, H. R., Wortmann, R. L., Saville, P. D., & Harkness, A. R. (1981). Hindsight bias among physicians weighing the likelihood of diagnoses. Journal of Applied Psychology, 66, Bartlett, F. C. (1932). Remembering; a study in experimental and social psychology. Oxford, England: Macmillan. Berlin, L. (1996a). Malpractice issues in radiology: Perceptual errors. American Journal of Roentgenology, 167, Berlin, L. (1996b). Errors in judgment. American Journal of Roentgenology, 166, Berlin, L. (2000). Malpractice issues in Radiology: Hindsight bias. American Journal of Roentgenology, 175, Berlin, L., & Hendrix, R. W. (1998). Malpractice issues in radiology: Perceptual errors and negligence. American Journal of Roentgenology, 170, Bernstein, D. M., Whittlesea, B. W. A., & Loftus, E. F. (2002). Increasing confidence in remote autobiographical memory and general knowledge: Extensions of the revelation effect. Memory and Cognition, 30, Bjorkman, M., Juslin, P., & Winman, A. (1993). Realism of confidence in sensory discrimination: The underconfidence phenomenon. Perception and Psychophysics, 54, Blackmore, S. J., Brelstaff, G., Nelson, K., & Troscianko, T. (1995). Is the richness of our visual world an illusion? Transsacadic memory for complex scenes. Perception, 24, Bradfield, A. L., Wells, G. L., & Olson, E. A. (2002). The damaging effect of confirming feedback on the relation between eyewitness certainty and identification accuracy. Journal of Applied Psychology, 87, Brainard, D. H. (1997). The Psychophysics Toolbox. Spatial Vision, 10,
22 HARLEY, CARLSEN, & LOFTUS 20 Bruner, J. S., & Potter, M. C. (1964). Interference in visual recognition. Science, 144, Chambres, P., Izaute, M., & Marescaux, P. (Eds.). (2002). Metacognition: Process, function and use. Boston, MA: Kluwer Academic Publishers. Christensen-Szalanski, J. J., & Willham, C. F. (1991). The hindsight bias: A meta-analysis. Organizational Behavior and Human Decision Processes, 48, Dawes, R. M. (1980). Confidence in intellectual vs. confidence in perceptual judgments. In E. D. Lanterman & H. Feger (Eds.), Similarity and choice: Papers in honour of Clyde Coombs (pp ). Bern, Switzerland: Huber. Fischhoff, B. (1975). Hindsight foresight: The effect of outcome knowledge on judgment under uncertainty. Journal of Experimental Psychology: Human Perception and Performance, 1, Fischhoff, B. (1977). Perceived informativeness of facts. Journal of Experimental Psychology: Human Perception and Performance, 3, Fischhoff, B. (1982). For those condemned to study the past: Heuristics in hindsight. In D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases. New York: Cambridge Univ. Press. Harley, E. M., Carlsen, K. A., & Loftus, G. R. (2001, November). Is hindsight 20/20?: Evidence for hindsight bias in visual perception. Poster session presented at the annual meeting of the Psychonomic Society, Orlando, FL. Hawkins, S. A., & Hastie, R. (1990). Hindsight: Biased judgments of past events after the outcomes are known. Psychological Bulletin, 107, Henderson, J.M. (1997). Transsacadic memory and integration during real-world object perception. Psychological Science, 8, Jacoby, L. L., & Whitehouse, K. (1989). An illusion of memory: False recognition influenced by unconscious perception. Journal of Experimental Psychology: General, 118, Keren, G. (1988). On the ability of monitoring non-veridical perceptions and uncertain knowledge: Some calibration studies. Acta Psychologica, 67,
PSYCHOLOGICAL SCIENCE. Research Report WE SAW IT ALL ALONG: VISUAL HINDSIGHT BIAS IN CHILDREN AND ADULTS
Research Report WE SAW IT ALL ALONG: VISUAL HINDSIGHT BIAS IN CHILDREN AND ADULTS Daniel M. Bernstein, Cristina Atance, Geoffrey R. Loftus and Andrew N. Meltzoff University of Washington, Seattle Abstract
More informationDoes scene context always facilitate retrieval of visual object representations?
Psychon Bull Rev (2011) 18:309 315 DOI 10.3758/s13423-010-0045-x Does scene context always facilitate retrieval of visual object representations? Ryoichi Nakashima & Kazuhiko Yokosawa Published online:
More informationHindsight Bias: A Cross-Cultural Analysis. STEVEN J. HEINE and DARRIN R. LEHMAN University of British Columbia, Canada
The Japanese Journal of Experimental Social Psychology. 1996, Vol. 35, No.3, 317-323 \ General Article \ Hindsight Bias: A Cross-Cultural Analysis STEVEN J. HEINE and DARRIN R. LEHMAN University of British
More informationHow Many Colors Can You Remember? Capacity is about Conscious vs unconscious memories
Science B44 Lecture 18 Visual Memory Memory 1. Afterimage, persistence, iconic sensory memory 2. Conscious vs unconscious memories 3. Short and long term memories 4. Where are memories seen 5. Flashbulb
More informationPsychological. Influences on Personal Probability. Chapter 17. Copyright 2005 Brooks/Cole, a division of Thomson Learning, Inc.
Psychological Chapter 17 Influences on Personal Probability Copyright 2005 Brooks/Cole, a division of Thomson Learning, Inc. 17.2 Equivalent Probabilities, Different Decisions Certainty Effect: people
More informationPerceptual Fluency Affects Categorization Decisions
Perceptual Fluency Affects Categorization Decisions Sarah J. Miles (smiles25@uwo.ca) and John Paul Minda (jpminda@uwo.ca) Department of Psychology The University of Western Ontario London, ON N6A 5C2 Abstract
More informationTHE RELIABILITY OF EYEWITNESS CONFIDENCE 1. Time to Exonerate Eyewitness Memory. John T. Wixted 1. Author Note
THE RELIABILITY OF EYEWITNESS CONFIDENCE 1 Time to Exonerate Eyewitness Memory John T. Wixted 1 1 University of California, San Diego Author Note John T. Wixted, Department of Psychology, University of
More informationMagnitude and accuracy differences between judgements of remembering and forgetting
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY 2012, 65 (11), 2231 2257 Magnitude and accuracy differences between judgements of remembering and forgetting Michael J. Serra and Benjamin D. England Department
More information(Visual) Attention. October 3, PSY Visual Attention 1
(Visual) Attention Perception and awareness of a visual object seems to involve attending to the object. Do we have to attend to an object to perceive it? Some tasks seem to proceed with little or no attention
More informationSatiation in name and face recognition
Memory & Cognition 2000, 28 (5), 783-788 Satiation in name and face recognition MICHAEL B. LEWIS and HADYN D. ELLIS Cardiff University, Cardiff, Wales Massive repetition of a word can lead to a loss of
More informationCognitive Psychology. Mark Rafter Multiple Intelligences
Cognitive Psychology Mark Rafter http://www.canyons.edu/faculty/rafterm Roster: Please put a checkmark next to your name or add your name. Multiple Intelligences Handouts: Please pick up a copy of today
More informationThe eyes fixate the optimal viewing position of task-irrelevant words
Psychonomic Bulletin & Review 2009, 16 (1), 57-61 doi:10.3758/pbr.16.1.57 The eyes fixate the optimal viewing position of task-irrelevant words DANIEL SMILEK, GRAYDEN J. F. SOLMAN, PETER MURAWSKI, AND
More informationSELECTIVE ATTENTION AND CONFIDENCE CALIBRATION
SELECTIVE ATTENTION AND CONFIDENCE CALIBRATION Jordan Schoenherr, Craig Leth-Steensen, and William M. Petrusic psychophysics.lab@gmail.com, craig_leth_steensen@carleton.ca, bpetrusi@carleton.ca Carleton
More information3. Read the study by Grant. Underline psychology key words and add them to your glossary. 4. Make detailed notes on the study
Getting ready to study Psychology: 1. Read the study by Loftus and Palmer. Underline psychology key words and look up what they mean. Get yourself a small exercise book and start to make a glossary. 2.
More informationThe Impact of Schemas on the Placement of Eyes While Drawing.
The Red River Psychology Journal PUBLISHED BY THE MSUM PSYCHOLOGY DEPARTMENT The Impact of Schemas on the Placement of Eyes While Drawing. Eloise M. Warren. Minnesota State University Moorhead Abstract.
More informationFluency Misattribution and Auditory Hindsight Bias
Memory & Cognition (in press, June 2018) Fluency Misattribution and Auditory Hindsight Bias Daniel M. Bernstein Ragav Kumar and Michael E. J. Masson Daniel J. Levitin Kwantlen Polytechnic University University
More informationBrad Schaffer Forensic Psychology July 22, Schaffer 1
Brad Schaffer Forensic Psychology July 22, 2004 Schaffer 1 Schaffer 2 Credibility of Intellectually Disabled Eyewitnesses Since its first court appearance in 1985, DNA evidence has exonerated a multitude
More informationWhy do Psychologists Perform Research?
PSY 102 1 PSY 102 Understanding and Thinking Critically About Psychological Research Thinking critically about research means knowing the right questions to ask to assess the validity or accuracy of a
More informationRecognition Memory, Familiarity, and Déjà vu Experiences Anne M. Cleary
CURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE Recognition Memory, Familiarity, and Déjà vu Experiences Anne M. Cleary Colorado State University ABSTRACT Déjà vu occurs when one feels as though a situation
More informationRunning head: FALSE MEMORY AND EYEWITNESS TESTIMONIAL Gomez 1
Running head: FALSE MEMORY AND EYEWITNESS TESTIMONIAL Gomez 1 The Link Between False Memory and Eyewitness Testimonial Marianna L. Gomez El Paso Community College Carrie A. Van Houdt FALSE MEMORY AND EYEWITNESS
More informationImage generation in a letter-classification task
Perception & Psychophysics 1976, Vol. 20 (3),215-219 Image generation in a letter-classification task THOMAS R. HERZOG Grand Valley State Colleges, Allandale, Michigan 49401 Subjects classified briefly
More informationLeft Handed Split Brain. Learning Objectives Topics
Left Handed Split Brain Case study V.J.: Gazzaniga, 1998 Left handed split brain patient Spoke out of left hemisphere Wrote out of right hemisphere Writing = independent from language systems Frey et al.
More informationFukuoka University of Education
Tomoko Sugimura sugitomo@fukuoka-edu.ac.jp Fukuoka University of Education 18 5 6 facial perception, gender-discrimination, young children Developmental studies have demonstrated that young children inaccurately
More informationConfidence and Memory Accuracy: Do we know what we think we know? Bennett Brocka. University of Iowa
Running Head: CONFIDENCE AND MEMORY ACCURACY 1 Confidence and Memory Accuracy: Do we know what we think we know? Bennett Brocka University of Iowa Running Head: CONFIDENCE AND MEMORY ACCURACY 2 Abstract
More informationThe Role of Feedback in Categorisation
The Role of in Categorisation Mark Suret (m.suret@psychol.cam.ac.uk) Department of Experimental Psychology; Downing Street Cambridge, CB2 3EB UK I.P.L. McLaren (iplm2@cus.cam.ac.uk) Department of Experimental
More informationMisleading Postevent Information and the Memory Impairment Hypothesis: Comment on Belli and Reply to Tversky and Tuchin
Journal of Experimental Psychology: General 1989, Vol. 118, No. 1,92-99 Copyright 1989 by the American Psychological Association, Im 0096-3445/89/S00.7 Misleading Postevent Information and the Memory Impairment
More informationChange Blindness. Change blindness is the inability to recognize changes to objects in the environment
Change Blindness Change blindness is the inability to recognize changes to objects in the environment after interruption, i.e. an object in the environment changes while an individual is looking away,
More informationMemory. Chapter 7 Outline. Human Memory: Basic Questions. Memory 10/2/ Prentice Hall 1. Chapter 7. How is pulled back out ( ) from memory?
Memory Chapter 7 Chapter 7 Outline Basic Processing Encoding Retrieval Systems of Memory Physiology of Memory Your Memory Human Memory: Basic Questions How does get into memory? How is information in memory?
More informationThe Psychology of Inductive Inference
The Psychology of Inductive Inference Psychology 355: Cognitive Psychology Instructor: John Miyamoto 05/24/2018: Lecture 09-4 Note: This Powerpoint presentation may contain macros that I wrote to help
More informationExamples of Feedback Comments: How to use them to improve your report writing. Example 1: Compare and contrast
Examples of Feedback Comments: How to use them to improve your report writing This document contains 4 examples of writing and feedback comments from Level 2A lab reports, and 4 steps to help you apply
More informationWhen Eyewitnesses Are Also Earwitnesses: Effects on Visual and Voice Identifications
Butler University Digital Commons @ Butler University Scholarship and Professional Work - LAS College of Liberal Arts & Sciences 1993 When Eyewitnesses Are Also Earwitnesses: Effects on Visual and Voice
More informationActor-Observer Bias One of the most established phenomenon in social psychology YOUR behavior is attributed to OTHER S behavior is attributed to
1 The Self in Social Psychology Disclaimer: there are many class demonstrations not included in the notes (to prevent ruining them); if you miss this lecture, you should get the notes from someone in the
More informationtesting for implicit bias
testing for implicit bias impartial jury? [i]t is by now clear that conscious and unconscious racism can affect the way white jurors perceive minority defendants and the facts presented at their trials,
More informationEYEWITNESS IDENTIFICATION. Mary F. Moriarty SPD Annual Conference 2015
EYEWITNESS IDENTIFICATION Mary F. Moriarty SPD Annual Conference 2015 mary.moriarty@hennepin.us The Case In 1984, a college student named Jennifer Thompson was raped in her apartment in Burlington, North
More informationThe Good Judgment Project: A Large Scale Test of Different Methods of Combining Expert Predictions
AAAI Technical Report FS-12-06 Machine Aggregation of Human Judgment The Good Judgment Project: A Large Scale Test of Different Methods of Combining Expert Predictions Lyle Ungar, Barb Mellors, Ville Satopää,
More informationAuditory hindsight bias: Fluency misattribution versus memory reconstruction
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/312951624 Auditory hindsight bias: Fluency misattribution versus memory reconstruction Article
More informationImproving Individual and Team Decisions Using Iconic Abstractions of Subjective Knowledge
2004 Command and Control Research and Technology Symposium Improving Individual and Team Decisions Using Iconic Abstractions of Subjective Knowledge Robert A. Fleming SPAWAR Systems Center Code 24402 53560
More informationDiscovering Statistics: Experimental Project
Discovering Statistics: Experimental Project Overview Inthislaboratoryprojectyouwillhavetothinkupanexperiment,designtheexperimentyourselfandthencollectand analyseyourowndata.thisprojectisintendedtofitinwithwhatyou
More informationElizabeth Loftus. Lost in the mall study 1992
Lost in the mall study 1992 -Subject was told 4 stories from his past, including one that was made up about him being lost in the mall at age 5. -Over the next two weeks, he jotted down whatever memories
More informationComparative Ignorance and the Ellsberg Paradox
The Journal of Risk and Uncertainty, 22:2; 129 139, 2001 2001 Kluwer Academic Publishers. Manufactured in The Netherlands. Comparative Ignorance and the Ellsberg Paradox CLARE CHUA CHOW National University
More informationSchool of Psychology. Professor Richard Kemp. Cognitive Bias: Psychology and Forensic Science
School of Psychology Professor Richard Kemp Cognitive Bias: Psychology and Forensic Science Professor Bryan Found Cognitive Bias: Psychology and Forensic Science What is Cognitive Bias and why does it
More informationVolatile visual representations: Failing to detect changes in recently processed information
Psychonomic Bulletin & Review 2002, 9 (4), 744-750 Volatile visual representations: Failing to detect changes in recently processed information MARK W. BECKER and HAROLD PASHLER University of California,
More informationChanging expectations about speed alters perceived motion direction
Current Biology, in press Supplemental Information: Changing expectations about speed alters perceived motion direction Grigorios Sotiropoulos, Aaron R. Seitz, and Peggy Seriès Supplemental Data Detailed
More informationWSC 2018 SCIENCE. Science of Memory
WSC 2018 SCIENCE Science of Memory Schema 101 A schema describes a pattern of thought or behavior that organizes categories of information and the relationships among them. It can also be described as
More informationWhy Does Similarity Correlate With Inductive Strength?
Why Does Similarity Correlate With Inductive Strength? Uri Hasson (uhasson@princeton.edu) Psychology Department, Princeton University Princeton, NJ 08540 USA Geoffrey P. Goodwin (ggoodwin@princeton.edu)
More informationOctober 2, Memory II. 8 The Human Amnesic Syndrome. 9 Recent/Remote Distinction. 11 Frontal/Executive Contributions to Memory
1 Memory II October 2, 2008 2 3 4 5 6 7 8 The Human Amnesic Syndrome Impaired new learning (anterograde amnesia), exacerbated by increasing retention delay Impaired recollection of events learned prior
More informationInductive reasoning in the courtroom: Judging guilt based on uncertain evidence
Inductive reasoning in the courtroom: Judging guilt based on uncertain evidence Ann M. Martin (Ann.Martin@unsw.edu.au) School of Psychology, University of New South Wales Sydney, 2052, Australia Brett
More informationEyewitness Evidence. Dawn McQuiston School of Social and Behavioral Sciences Arizona State University
Eyewitness Evidence Dawn McQuiston School of Social and Behavioral Sciences Arizona State University Forensic Science Training for Capital Defense Attorneys May 21, 2012 My background Ph.D. in Experimental
More informationChapter 1 Observation Skills
Observation vs. Inference Chapter 1 Observation Skills Forensic relating to the application of scientific knowledge to legal questions must have ability to observe, interpret, and report observations clearly.
More informationAQA A Level Psychology. Topic WORKSHEETS. Memory.
AQA A Level Psychology Topic WORKSHEETS Memory Page 2 AQA A LEVEL Psychology topic worksheets: MEMORY THE MULTI STORE MODEL Specification: The multi store model of memory: sensory register, short term
More informationConfirmation Bias. this entry appeared in pp of in M. Kattan (Ed.), The Encyclopedia of Medical Decision Making.
Confirmation Bias Jonathan D Nelson^ and Craig R M McKenzie + this entry appeared in pp. 167-171 of in M. Kattan (Ed.), The Encyclopedia of Medical Decision Making. London, UK: Sage the full Encyclopedia
More informationI know your face but not where I saw you: Context memory is impaired for other-race faces
Psychonomic Bulletin & Review 2008, 15 (3), 610-614 doi: 10.3758/PBR.15.3.610 I know your face but not where I saw you: Context memory is impaired for other-race faces RUTH HORRY University of Sussex,
More informationMemory 2/15/2017. The Three Systems Model of Memory. Process by which one encodes, stores, and retrieves information
Chapter 6: Memory Memory Process by which one encodes, stores, and retrieves information The Three Systems Model of Memory Each system differs in terms of span and duration 1 The Three Systems Model of
More informationWendy J. Reece (INEEL) Leroy J. Matthews (ISU) Linda K. Burggraff (ISU)
INEEL/CON-98-01115 PREPRINT Estimating Production Potentials: Expert Bias in Applied Decision Making Wendy J. Reece (INEEL) Leroy J. Matthews (ISU) Linda K. Burggraff (ISU) October 28, 1998 October 30,
More informationThe Meaning of the Mask Matters
PSYCHOLOGICAL SCIENCE Research Report The Meaning of the Mask Matters Evidence of Conceptual Interference in the Attentional Blink Paul E. Dux and Veronika Coltheart Macquarie Centre for Cognitive Science,
More informationThe Role of Color and Attention in Fast Natural Scene Recognition
Color and Fast Scene Recognition 1 The Role of Color and Attention in Fast Natural Scene Recognition Angela Chapman Department of Cognitive and Neural Systems Boston University 677 Beacon St. Boston, MA
More informationMisled Subjects May Know More Than Their Performance Implies
Journal of Experimental Psychology: Learning, Memory, and Cognition 1989, Vol. 15, No. 2, 246-255 Copyright 1989 by the American Psychological Association, Inc. O278-7393/89/S0O.75 Subjects May Know More
More informationDoes Working Memory Load Lead to Greater Impulsivity? Commentary on Hinson, Jameson, and Whitney (2003)
Journal of Experimental Psychology: Learning, Memory, and Cognition 2006, Vol. 32, No. 2, 443 447 Copyright 2006 by the American Psychological Association 0278-7393/06/$12.00 DOI: 10.1037/0278-7393.32.2.443
More informationCategorization vs. Inference: Shift in Attention or in Representation?
Categorization vs. Inference: Shift in Attention or in Representation? Håkan Nilsson (hakan.nilsson@psyk.uu.se) Department of Psychology, Uppsala University SE-741 42, Uppsala, Sweden Henrik Olsson (henrik.olsson@psyk.uu.se)
More informationLimitations of Object-Based Feature Encoding in Visual Short-Term Memory
Journal of Experimental Psychology: Human Perception and Performance 2002, Vol. 28, No. 2, 458 468 Copyright 2002 by the American Psychological Association, Inc. 0096-1523/02/$5.00 DOI: 10.1037//0096-1523.28.2.458
More informationCHAPTER 3: SOCIAL PERCEPTION: UNDERSTANDING OTHER PEOPLE CHAPTER OVERVIEW
CHAPTER 3: SOCIAL PERCEPTION: UNDERSTANDING OTHER PEOPLE CHAPTER OVERVIEW Chapter 3 covers three main topics: impressions others make on us, attribution processes and biases, and the impressions we make
More informationAssessing the influence of recollection and familiarity in memory for own- vs. other-race faces
Iowa State University From the SelectedWorks of Christian A. Meissner, Ph.D. 2009 Assessing the influence of recollection and familiarity in memory for own- vs. other-race faces Jessica L Marcon, University
More informationCalibration in tennis: The role of feedback and expertise
Calibration in tennis: The role of feedback and expertise Gerard J. Fogarty (fogarty@usq.edu.au) Department of Psychology University of Southern Queensland, Toowoomba QLD 4350 Australia Anthony Ross (a_ross4@hotmail.com)
More informationObject Substitution Masking: When does Mask Preview work?
Object Substitution Masking: When does Mask Preview work? Stephen W. H. Lim (psylwhs@nus.edu.sg) Department of Psychology, National University of Singapore, Block AS6, 11 Law Link, Singapore 117570 Chua
More informationThe Effects of Action on Perception. Andriana Tesoro. California State University, Long Beach
ACTION ON PERCEPTION 1 The Effects of Action on Perception Andriana Tesoro California State University, Long Beach ACTION ON PERCEPTION 2 The Effects of Action on Perception Perception is a process that
More informationExploring a Counterintuitive Finding with Methodological Implications
Exploring a Counterintuitive Finding with Methodological Implications Why is 9 > 221 in a Between-subjects Design? Stuart J. McKelvie Department of Psychology Bishop s University 2600 College Street, Sherbrooke,
More informationSubliminal Programming
Subliminal Programming Directions for Use Common Questions Background Information Session Overview These sessions are a highly advanced blend of several mind development technologies. Your mind will be
More informationA Critical Mass of Knowledge: Heightened Stress and Retention Interval
A Critical Mass of Knowledge: Heightened Stress and Retention Interval Estimator variables: Criteria for a critical mass of knowledge 1. There should have been published at least one up-to-date meta-analysis
More informationDecisions based on verbal probabilities: Decision bias or decision by belief sampling?
Decisions based on verbal probabilities: Decision bias or decision by belief sampling? Hidehito Honda (hitohonda.02@gmail.com) Graduate School of Arts and Sciences, The University of Tokyo 3-8-1, Komaba,
More informationSupporting Information
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Supporting Information Variances and biases of absolute distributions were larger in the 2-line
More informationJournal of Experimental Social Psychology
Journal of Experimental Social Psychology 45 (2009) 1117 1122 Contents lists available at ScienceDirect Journal of Experimental Social Psychology journal homepage: www.elsevier.com/locate/jesp Reports
More informationChapter 11. Experimental Design: One-Way Independent Samples Design
11-1 Chapter 11. Experimental Design: One-Way Independent Samples Design Advantages and Limitations Comparing Two Groups Comparing t Test to ANOVA Independent Samples t Test Independent Samples ANOVA Comparing
More informationDetecting and identifying response-compatible stimuli
Psychonomic Bulletin & Review 1997, 4 (1), 125-129 Detecting and identifying response-compatible stimuli JOCHEN MÜSSELER and BERNHARD HOMMEL Max Planck Institute for Psychological Research, Munich, Germany
More informationAttentional Theory Is a Viable Explanation of the Inverse Base Rate Effect: A Reply to Winman, Wennerholm, and Juslin (2003)
Journal of Experimental Psychology: Learning, Memory, and Cognition 2003, Vol. 29, No. 6, 1396 1400 Copyright 2003 by the American Psychological Association, Inc. 0278-7393/03/$12.00 DOI: 10.1037/0278-7393.29.6.1396
More informationSelective attention and asymmetry in the Müller-Lyer illusion
Psychonomic Bulletin & Review 2004, 11 (5), 916-920 Selective attention and asymmetry in the Müller-Lyer illusion JOHN PREDEBON University of Sydney, Sydney, New South Wales, Australia Two experiments
More informationEyewitness Testimony. Student s Name. Institution of Learning
Running head: EYEWITNESS TESTIMONY 1 Eyewitness Testimony Student s Name Institution of Learning EYEWITNESS TESTIMONY 2 In a legal proceeding, evidence serves as a critical component of reaching a decision.
More informationA model of parallel time estimation
A model of parallel time estimation Hedderik van Rijn 1 and Niels Taatgen 1,2 1 Department of Artificial Intelligence, University of Groningen Grote Kruisstraat 2/1, 9712 TS Groningen 2 Department of Psychology,
More informationPART 1. Have I SEEN that before? Theories of recognition memory. Structure of talk. Memory strength theory. Signal detection theory.
Have I SEEN that before? Theories of recognition memory Structure of talk Memory strength theory Prof Tim Perfect Challenges to memory strength theory Dual-process theories RK theory Attributional theories
More informationIntentional and Incidental Classification Learning in Category Use
Intentional and Incidental Classification Learning in Category Use Michael Romano (mrr2@nyu.edu) Department of Psychology, New York University, 6 Washington Place New York, NY 1000 USA Abstract Traditional
More information20. Experiments. November 7,
20. Experiments November 7, 2015 1 Experiments are motivated by our desire to know causation combined with the fact that we typically only have correlations. The cause of a correlation may be the two variables
More informationAssessing reliability
Assessing reliability Leslie Cuthbert Recorder, Tribunal Judge & Adjudicator United Kingdom Session Aims There are 3 areas that will be covered in this session: What is the difference between credibility
More informationWhat is schema theory?
Schema Theory What is schema theory? The term schema was first used by Jean Piaget in 1926. (but there have been many with similar ideas before him) Explains how people develop ideas from simple ideas
More informationNon-categorical approaches to property induction with uncertain categories
Non-categorical approaches to property induction with uncertain categories Christopher Papadopoulos (Cpapadopoulos@psy.unsw.edu.au) Brett K. Hayes (B.Hayes@unsw.edu.au) Ben R. Newell (Ben.Newell@unsw.edu.au)
More informationForced confabulation affects memory sensitivity as well as response bias
Mem Cogn (2012) 40:127 134 DOI 10.3758/s13421-011-0129-5 Forced confabulation affects memory sensitivity as well as response bias Victor Gombos & Kathy Pezdek & Kelly Haymond Published online: 23 July
More informationEveryone asks and answers questions of causation.
Chapter 11: Cause and Effect: Explaining Why Everyone asks and answers questions of causation. Cause and effect, like the two sides of a coin, are inseparably linked and together make up causation. Cause
More informationPartial Representations of Scenes in Change Blindness: In the Eyes and in the Hands
Partial Representations of Scenes in Change Blindness: In the Eyes and in the Hands Nick Benesh (nbenesh@memphis.edu) Brent Fonville (jbfnvlle@memphis.edu) Rick Dale (radale@memphis.edu) Department of
More informationFraming the frame: How task goals determine the likelihood and direction of framing effects
McElroy, T., Seta, J. J. (2007). Framing the frame: How task goals determine the likelihood and direction of framing effects. Judgment and Decision Making, 2(4): 251-256. (Aug 2007) Published by the Society
More informationTIME-ORDER EFFECTS FOR AESTHETIC PREFERENCE
TIME-ORDER EFFECTS FOR AESTHETIC PREFERENCE Åke Hellström Department of Psychology, Stockholm University, S-106 91 Stockholm, Sweden Email: hellst@psychology.su.se Abstract Participants compared successive
More informationMeta-Analyses of Estimator and System Variables
Meta-Analyses of Estimator and System Variables Meta-analyses of Estimator and System variables A quantitative review, combining tests of a common hypothesis Summarizes reliable patterns of outcomes, across
More informationExam 2 PS 306, Spring 2004
Exam 2 PS 306, Spring 2004 1. Briefly define the term confound. Then, using a very explicit example of practice effects (maybe even with numbers?), illustrate why conducting a repeated measures experiment
More informationAccess to justice for child witnesses on the autism spectrum. Lucy Henry, Laura Crane and Rachel Wilcock
Access to justice for child witnesses on the autism spectrum Lucy Henry, Laura Crane and Rachel Wilcock 1 Background to the project A small but growing literature indicates that children on the autism
More informationImplicit Information in Directionality of Verbal Probability Expressions
Implicit Information in Directionality of Verbal Probability Expressions Hidehito Honda (hito@ky.hum.titech.ac.jp) Kimihiko Yamagishi (kimihiko@ky.hum.titech.ac.jp) Graduate School of Decision Science
More informationProject exam in Cognitive Psychology PSY1002. Autumn Course responsible: Kjellrun Englund
Project exam in Cognitive Psychology PSY1002 Autumn 2007 674107 Course responsible: Kjellrun Englund Stroop Effect Dual processing causing selective attention. 674107 November 26, 2007 Abstract This document
More informationOlder adults associative deficit in episodic memory: Assessing the role of decline in attentional resources
Psychonomic Bulletin & Review 2004, 11 (6), 1067-1073 Older adults associative deficit in episodic memory: Assessing the role of decline in attentional resources MOSHE NAVEH-BENJAMIN University of Missouri,
More informationFAQ: Heuristics, Biases, and Alternatives
Question 1: What is meant by the phrase biases in judgment heuristics? Response: A bias is a predisposition to think or act in a certain way based on past experience or values (Bazerman, 2006). The term
More informationMemory Schemas, Source Monitoring & Eyewitness Memory
Memory Schemas, Source Monitoring & Eyewitness Memory Psychology 355: Cognitive Psychology Instructor: John Miyamoto 05/09/2018: Lecture 07-3 Note: This Powerpoint presentation may contain macros that
More informationo^ &&cvi AL Perceptual and Motor Skills, 1965, 20, Southern Universities Press 1965
Ml 3 Hi o^ &&cvi AL 44755 Perceptual and Motor Skills, 1965, 20, 311-316. Southern Universities Press 1965 m CONFIDENCE RATINGS AND LEVEL OF PERFORMANCE ON A JUDGMENTAL TASK 1 RAYMOND S. NICKERSON AND
More informationThe Instrumentality Heuristic
PSYCHOLOGICAL SCIENCE Research Article The Instrumentality Heuristic Why Metacognitive Difficulty Is Desirable During Goal Pursuit Aparna A. Labroo and Sara Kim University of Chicago ABSTRACT The literature
More informationHow much can you trust your memory?
How much can you trust your memory? How Much Do You Know? How much emphasis does the legal system place on eyewitness testimony? What factors can limit an eyewitness accuracy? What role do police officers
More informationEFFECTS OF NOISY DISTRACTORS AND STIMULUS REDUNDANCY ON VISUAL SEARCH. Laurence D. Smith University of Maine
EFFECTS OF NOISY DISTRACTORS AND STIMULUS REDUNDANCY ON VISUAL SEARCH Laurence D. Smith University of Maine ldsmith@maine.maine.edu Ronald M. Pickett Marjan Trutschl Institute for Visualization and Perception
More information